Practical Algorithms for Image Analysis: Description, Examples, and Code

Citation

Andrew, A.M. (2001), "Practical Algorithms for Image Analysis: Description, Examples, and Code", Kybernetes, Vol. 30 No. 1, pp. 103-115. https://doi.org/10.1108/k.2001.30.1.103.7

A very impressive range of practical algorithms is provided, in what is described in the book’s foreword as “classic cookbook style”. Image analysis finds application in medical imaging and in processing of photographs from aeroplanes and satellites and surveillance cameras. It is also applicable in a variety of other scientific contexts, and obviously in AI and robotics projects depending on visual input.

The authors adopt a business‐like style and do not waste time on general discussion of areas of application. Their biographical notes give some hints, though, and two of them have been involved in research on pattern formation in thin films of organic or magnetic material. The first author is president of a high‐tech company developing systems for medical diagnostics, and the second is in a company developing biometric authentication systems for personal access, a topic currently receiving much attention.

This is undoubtedly an area where it is useful to have well‐tried algorithms available. As well as the assurance that the algorithms do what they should, the time‐efficiency of operations on two‐dimensional data can depend enormously on how the processing is organised. This last point can become evident even in the simple context of the cellular automaton simulations and the game of “Life” as described by Gardner (1970; 1971).

The programs on the CD‐ROM are in the C programming language and are conveniently labelled according to the sections of the book in which they are described. A feature of C is that algorithms can conveniently be chained so that the output of one becomes the input of the next.

The programs will run on a Windows (95, 98 or NT) or Linux platform and require a C or C++ compiler.

Most people who use the book will be already active in C programming, but it is interesting to note that a C compiler can be downloaded from the site: <http://www.cs.virginia.edu/‐lcc‐win32>. The “readme” file on the CD‐ROM mentions that the development of the programs was done using Microsoft Visual C++, version 40. The images have to be in uncompressed TIFF format, and the CD‐ROM gives advice on arranging this using packages such as Adobe Photoshop.

The first chapter of the book gives a useful overview of the remainder, and a number of theoretical aspects are treated in an Appendix. Between these, Chapter 2 is on “Global image analysis”, with attention to intensity histograms, global enhancement and geometric image transformations, Chapter 3 is on “Grey‐scale image analysis” with attention to local image operations and noise reduction, as well as edge detection and template matching, and Chapter 4 is on “Binary image analysis”, with attention to morphological and cellular processing, and various operations referring to shape analysis, including determination of convex hulls and inference of straight lines using the Hough transform.

Chapter 5 is on “Analysis of lines and line patterns” and includes polygonalization as well as straight‐line and cubic spline fitting, while Chapter 6 is on “Analysis of point patterns”, including reference to Voronoi diagrams, and Chapter 7 is on “Frequency domain analysis”. This listing of topics is by no means complete but should give some idea of the wide coverage of this book. With its CD‐ROM, it is likely to come to be accepted as virtually indispensable for the efficient setting up of projects involving image analysis.